Unwinding A Large Position Gracefully

You built the position slowly, over days, buying into a name you believed in. Now the thesis has played out, or died, and you need to sell 400,000 shares of a stock that trades $18m a day. The decision to exit is the easy part. How you exit decides whether you keep the profit you think you have.

The trap is treating the unwind as an afterthought — "the model says sell, so sell" — when the mechanics of how you sell 400,000 shares can cost more than the difference between a good entry and a mediocre one. A position is not really worth its last mark until you know what it is worth on the way out, and those are two different numbers.

Why the exit is harder than the entry

When you built the position, the market didn't know your final size — you could buy patiently, in pieces, disguised in the day's normal flow. On the way out, the same thing is true in reverse, except now there is a second problem: the market may already suspect you're exiting, because your entry told other participants there's a large holder to unwind eventually. A slow, obvious seller invites other people to sell in front of them.

Two costs work against you simultaneously. Sell too fast and you push the price down against yourself — market impact. Sell too slow and you sit exposed to the market moving against you for longer — timing risk. Every unwind is a trade-off between these two, and the right balance depends on how urgent your reason for leaving is.

The main choices

Participation rate. Selling at a fixed percentage of volume — say 12 percent of every print — spreads the order across the day and scales naturally with liquidity: busy periods get more of your size, quiet periods less.

Time-based schedule (VWAP/TWAP). Selling a fixed amount per interval regardless of volume. Simpler, more predictable, worse in a thin or volatile tape because it doesn't adapt.

Opportunistic blocks. Waiting for a natural buyer — a cross, a block desk, an ETF creation flow — and transacting size in one print instead of many. Best price per share when it works; you might wait days for a print that never comes.

Dark pools and algos. Routing through venues designed to hide size, at the cost of less certainty about when you'll actually get filled.

None of these is universally right. A position you must be out of by Friday because of a redemption cannot wait for the perfect block. A position you're trimming because it got a little too big for the concentration cap can wait weeks for a good print.

Worked example

You hold 400,000 shares of a stock last marked at $52.10, average daily volume $18m (roughly 350,000 shares). Your desk head wants the position materially reduced within the week because the thesis has weakened and correlation to the rest of the book has crept up.

You choose 15 percent of volume as your participation rate: roughly 52,500 shares a day, meaning the unwind takes about eight sessions. On day one you fill 51,000 shares at an average of $52.02, eight cents below the prior mark — normal impact for that size in that name. By day four the stock has drifted to $51.40 on its own (unrelated news, a soft sector day), and your day-four fills average $51.30. You are not fighting your own selling at that point; you are fighting the market, which is exactly why the schedule existed — it never bet the whole exit on one day's tape.

Compare that to the alternative: dumping all 400,000 shares in the first ninety minutes. On a $18m ADV name that is over four times a typical hour's volume; realistic estimates put the impact at 3 to 5 percent, meaning fills averaging closer to $50 — over a million dollars worse than the patient schedule, for a position that had no reason to be that urgent.